A barcode shape descriptor for curve point cloud data

In this paper, we present a complete computational pipeline for extracting a compact shape descriptor for curve point cloud data. Our shape descriptor, called a barcode, is based on a blend of techniques from differential geometry and algebraic topology. We also provide a metric over the space of barcodes, enabling fast comparison of PCDs for shape recognition and clustering. To demonstrate the feasibility of our approach, we have implemented it and provide experimental evidence in shape classification and parametrization.

• Type: Journal
• Year: 2004
• Where: CG
• Authors: Anne D. Collins, Afra Zomorodian, Gunnar Carlsson, Leonidas J. Guibas
• Keywords: CG 2004, Compact Shape Descriptor, Complete Computational Pipeline, Computer Graphics, Shape Descriptor